Re: libX11's libXcursor dlopen

> dlopens() "libXcursor.so".  note the lack of version.  [...]

> on my system however I see this:

> 24472  24472 example  CALL  open(0x7f7fff2a4618,0,6)
> 24472  24472 example  NAMI  "/usr/X11R7/lib/libXcursor.so.2"
> 24472  24472 example  RET   open 4

> where does that query for version '2' of the library come from?

I'm not sure (at a quick glance I'm having trouble finding the source
to dlopen()).  My first guess is that it comes from the library major
version the binary was linked with, but I'm unsure enough of that that
I would not call it anything more than a guess.  It is, though,
supported by link(5) on a 9.1 machine at work, which describes a struct
sod, which "describes a shared object that is needed to complete the
link edit process of the object containing it" and includes integer
major and minor version numbers as well as a library name.

> 2. is it normal for a NetBSD system to not even have a /etc/ld.so.conf file?$

That 9.1 machine at work has none and appears to be fine from what I
can tell, including running X (but not specifically using libXcursor).

> 4. where can I find the sources for that elusive libXcursor.so.2 that comes $

I don't know.  Have you tried doing a full build of the world?  That's
one of the first things I do with a new system, and one reason is that
that's the only way I know of I can be certain I have the source that
corresponds to the binaries I'm running.  (You don't have to install
the resultant build, though if you do and it changes anything then
something's broken somewhere in NetBSD's release process, in that it
would mean they're shipping binaries and sources that don't match.)

The build log from a build-of-the-world should allow you to determine
where the source any particular piece got built from is.  It certainly
does in the versions I use.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B